INDEX

Explanations

still

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 much

-0.53

 really

-0.48

much

-0.47

 some

-0.45

 still

-0.42

 textStatus

-0.41

however

-0.40

 idées

-0.40

really

-0.40

now

-0.39

POSITIVE LOGITS

parsedMessage

1.06

ScopeManager

0.81

jooq

0.78

ThroughAttribute

0.78

 AssemblyCulture

0.77

 незавершена

0.77

 '\\;'

0.76

ⓧ

0.76

 utafitiHapana

0.74

 noqa

0.73

Activations Density 0.114%